The Mitchell Line: Clearing the RB Bar
How can we define—and predict—running back prospect success?
I’ve spent much of the past few months digging through years of NFL Draft & fantasy football data, trying to uncover useful trends. In the process, I’ve found some success comparing the league’s Next Gen Stats (NGS) data with historical dynasty-league average draft position (ADP), examining both running backs and receivers in earlier posts.
What’s been a harder nut to crack, however, is predicting which late-round dynasty picks will succeed. In fact, it’s pretty much the entire challenge, given how high the hit rate is for first-rounders. Even if you whiff on your first-round pick, they’ll more often than not provide value by virtue of actually seeing the field.
Conversely, players picked after the first-round of dynasty drafts are significantly harder to peg. A metric that may be great at predicting success for second-rounders, for example, may flop at predicting third-rounders, underscoring just how noisy prospect evaluation can be. Yet this is something for later in the article; first we much define what prospect success actually looks like.
The Mitchell Line
Enter the Mitchell Line, a semi-arbitrary cutoff I’ve devised to judge whether a running back’s first three years in the league were—at least in fantasy terms—a success. In short, we’re looking for starters, or players who cleared 250 PPR points across their first three years in the league.
It’s named after former 49ers running back Elijah Mitchell, who overcame his late-round pedigree to unseat a player picked ahead of him in the same draft. Granted, he was meaningfully better than what I’d consider a typical starter-level player, with his career being derailed by injuries and the 49ers’ addition of Christian McCaffrey. Still, the line's more about the guys who don’t clear it than anything.
Dynasty managers were likely fine getting Jerome Ford and Khalil Herbert in the fourth round, for example. Those who picked Benny Snell or Mark Walton, however, were less happy, despite getting spurts of production and spot-start upside. They didn’t get a startable player, which is the point the Mitchell line tries to convey: you’re here to draft starters, not fringe backs who’ll be off your roster in a year.
The story told by the graph above—which compares backs above and below the Mitchell line by athleticism—is more about how un-athletic the below-line backs are than anything. The worst back above it is Kyren Williams in dark blue, with an NGS Athleticism Score (more info here) of 60, the only blue dot above the line. Despite boasting some strong athletes, players below the line look more like Williams than they do Mitchell and his blindingly yellow 99 athletic score (out of 100), a mix of ho-hum purples and blues that suggest avoiding subpar athletes in the back half of dynasty drafts.
Comparing RB Attributes by Round
While interesting enough, our athleticism analysis only tells a small part of the story, given we were only looking at dynasty picks 30-plus. This was deliberate, however, since that slice encompasses one of our biggest blind spots: fourth-round picks. There’s little difference, for example, in the NGS Score (rightmost group) of successful and unsuccessful fourth-round prospects.
It’s also harder to differentiate good from bad first-rounders, though their high hit rate makes that issue somewhat moot. Ultimately, there aren’t many one-size-fits-all metrics that help distinguish successful prospects from busts across all rounds of dynasty drafts. The best angle, in my opinion, is finding criteria like the ones above (many sourced from Peter Howdy’s excellent spreadsheet), where successful backs hold an advantage in across all four rounds of dynasty drafts.
Going beyond success & looking for stars
Beyond using round-based ratios, I also looked at correlations between our metrics and whether a player cleared the Mitchell line (the blue bars above). I also wanted to better suss out stars from just-OK players, so for each indicator, I compared its correlation with early-career production for players already above the Mitchell line (red bars). The idea here is that if we can find an attribute that’s specifically predictive for players we already think are going to succeed, we might be able to go a step beyond and find not just starters, but stars.
For example, Overall NGS Score (leftmost bars) was the feature that best predicted whether a player would surpass the Mitchell line, boasting a correlation coefficient of .48. When we looked solely at players above that line, the correlation with total points scored was still a comparatively robust .49. Bottom line: we want to find metrics like NGS Score, which not only predict player success, but help explain just how successful a good prospect will be.
2025 RB class breakdown
OK, great, you’re probably thinking; but what about the actual prospects? While the 2025 RB class looks to be one of the better ones in recent memory, there’s still only one elite-level prospect in Ashton Jeanty. It is worth noting that, in terms of NGS score, most drafts are lucky to have one prospect at Jeanty’s level, with many having only a single player grading out above 80. Here, five players clear that threshold, with others close behind.
What I find under-discussed about the 2025 class, however, is how few backs are the complete package. In a way, the league’s draft approach bears this out, with many teams selecting backs to compliment their existing backs’ skillsets. Take, for example, the Giants selecting Cam Skattebo as a steady-hand counterpart to the explosive Tyrone Tracy. Also see the Browns, who double-dipped and got both flavors in the sturdy Quinshon Judkins and the more dynamic Dylan Sampson (who are, quite fittingly, neighbors on the chart above).
Upper-flight talents
All this is to say, while I’d usually advocate for doing RB analysis independent of landing spot, I think that this year, team-fit concerns matter a lot more than usual. Take, for example, RJ Harvey, a player I’d have happily acquired for a third-round pick, which was his going rate before NFL Draft. By our metrics, he’s an upper-quadrant player, a productive receiving threat with solid athletic traits.
Now that Harvey’s likely to command a first-round selection, though, things have changed. Much of his analytical upside is already baked into that price, and his shortcomings now stand out more sharply. Harvey is, as Nate Tice described Oregon QB Dillon Gabriel, “old and short,” which raises questions about his ability to sustain a full workload
Many other high-profile players in the chart above, like Harvey, offer little surplus value at their current ADP: you don’t need me to tell you TreyVeon Henderson is going to be good, for example. The edge instead comes from guys like Bhayshul Tuten, an exceptional receiver with explosive traits. He’s a great bet to succeed in Liam Coen’s offense, who made a three-headed monster work in Tampa last year.
Tuten is starting to run into the same issue as Harvey, however, offering less value the earlier his ADP gets. Consider, then, DJ Giddens, whose receiving upside should let him see the field early, and scores comparably to Tuten in most of our metrics. Perhaps an even better value bet is Ollie Gordon: while Gordon’s metrics are inflated by an outlier 2023 season, he boasts a degree of size that Miami sorely lacked.
Sifting through the rest
As for the rest of the 2025 class? Decidedly “meh” by comparison. More importantly, these backs all rank similarly in age-adjusted PPR per Game (our y-axis here), which is an especially strong predictor of success for third-round prospects. Though many of these guys are actually fourth-rounders, the stat’s still useful—as is Average Yards per Route Run (our new x-axis)—in determining prospect quality.
Since most players above are ranked similarly, we should first hone in on outliers. Jordan James stands out for his particularly poor age-adjusted production, though it’s worth noting he had to share a backfield with Bucky Irving at Oregon. Even if he wins Kyle Shanahan’s favor, however, he’s still decently unlikely to eclipse our 250-PPR benchmark, with his middling receiving profile likely capping his upside.
Maybe, then, we should turn to Jaydon Blue, the best receiving back in the chart above, who also boasts blazing speed. Woody Marks stands out, too, with an enticing blend of pedigree (fourth-round NFL Draft pick) and receiving ability that should give him a chance to climb up a weak Texans depth chart. The difference, of course, is that Blue’s continued rise up draft boards makes him less of a value pick, while it’s hard to see Marks’ ADP changing further.
Simple thresholds for ‘25 RBs
Since differentiating between these late-round backs is so difficult, let’s once again turn to NGS Athleticism & Production Scores, using a score of 65 as a rough cutoff for success. Only twos backs fail to clear our production threshold: Trevor Etienne and Woody Marks. This is a considerable red flag; as I’ve covered in a previous post, Production Score is arguably our best metric at predicting RB success.
Less disconcerting—though still notable—are our three athletic outliers, Devin Neal, Kaleb Johnson and Ollie Gordon, whom we covered earlier. If anything, Neal’s low athleticism only serves to highlight how strong his Production Score (seventh-best in the class) is relative to other backs with a third-round ADP. That makes him an interesting sleeper, especially with Alvin Kamara getting up there in years.
Johnson’s a slightly more interesting case, if only because his poor combine results overshadow his superb tracking numbers. If he is, in actuality, a stronger athlete than his testing shows, then the only thing holding him back would be his weak receiving profile. While there are still enough red flags to make me wary, the further he falls down draft boards, the more willing I am to bet on his three-down-back potential.
It all boils down to opportunity cost, which is a good lens to reexamine our low-production outliers through. Etienne in the third, for example, feels slightly rich, given how little he’s done in college. By contrast, Woody Marks is a worthy dart-throw in the fourth due to his receiving ability and Joe Mixon’s advanced age, even if it means engaging in Tobias Fünke logic regarding his suspect Production Score.
Summary
Let’s codify what good RB prospects should do:
Clear the bar for both production & athleticism
Be more than their best season (age-adjusted PPR/G)
Be holistically good (Overall NGS)
Be competent receivers (Avg. Y/RR)
There are a couple more concepts that fit less cleanly into bullet points. First, backs should be picked appropriately based on NFL Draft pedigree. Landing spot is a great for differentiating players of comparable pedigree, but that doesn’t mean you should select Isaiah Spiller (the 123rd overall pick!) in the early second just because of a good landing spot.
Second, age and consistency matters when it comes to production. Any player that had a single breakout season—or played against weaker competition—should be met with scrutiny. This is doubly true if they did so as an older prospect, for both wear-and-tear and credibility reasons; an older dude with more experience will obviously be better equipped to mow down freshmen. This is the Cam Skattebo conundrum, one with no easy answers, but at least is enough to arm us with a health dose of skepticism.
Yet if there’s one idea I want everyone reading this to grasp, it’s this: predicting prospect success is still hard. In a lot of senses, it’s meaningfully easier to suss out which RB’s will do well, especially given how highly college production correlates with future success. But it’s still noisy enough that the real game shouldn’t about who you think will be good, but what price you’re willing to pay for a given prospect.
Hopefully, all this has equipped you with enough tools to find some winners in your dynasty league drafts. I once again want to thank the illustrious Peter Howdy, who you can find here on Bluesky. You can similarly find me on Twitter (capn_cc) and Bluesky, and feel free to subscribe to this Substack (free!) for more in-depth fantasy content.